Marco Fratus

Wearable glucose monitors and insulin patches have quickly become a familiar sight in TV commercials and on social media, offering people with diabetes — and increasingly those using weight-loss medications — a discreet way to track and manage blood sugar. But while the devices look simple on the outside, the science that connects the human body’s glucose-insulin cycle to wearable technology has remained a “black box.”

A new study from Purdue University’s Elmore Family School of Electrical and Computer Engineering closes that gap by introducing a predictive, closed-loop model that could shape the future of personalized diabetes care.

Published in Proceedings of the National Academy of Sciences, the research by former PhD student Marco Fratus and Muhammad A. Alam, the Jai N. Gupta Distinguished Professor of ECE, introduces a first-of-its-kind mathematical framework linking body to device operation, and vice versa. In doing so, this framework treats the human body and the wearable patch as a single integrated system, linking how the body naturally regulates glucose to the specific physics of how microneedle patches sense and deliver insulin, and offering a tool to accelerate innovations in diabetes management.

“This work finally connects human physiology with device physics in a common mathematical framework,” Alam said. “It allows us to move beyond observing clinical trials to predicting outcomes. We can now ask and answer new questions about how a patch’s physical design should be personalized to match an individual's unique metabolic signature.”

A blueprint for a smarter artificial pancreas

Today’s automated insulin delivery systems, sometimes described as an “artificial pancreas,” combine continuous glucose monitors with insulin pumps. But progress toward a fully automated, closed-loop system has been slowed by trial-and-error device development and a limited understanding of how design decisions affect the body’s real-time glucose control.

The Purdue model ties those worlds together. It predicts how microneedle-patch design choices like needle size, shape, and material influence how quickly and accurately glucose is sensed and how insulin is delivered into the body.  PNAS reviewers called the work “timely” and “elegant.”

New possibilities: early disease signals and smarter eating habits

Because wearable sensors continuously record glucose and insulin levels, the research raises intriguing new clinical possibilities. Could glucose-insulin oscillations — the natural rise and fall of sugar and insulin — become a kind of “EKG” for pancreas and kidney health? The model suggests that deviations from these predicted rhythms could serve as early warning signals for organ dysfunction.

And in a nod to food-science curiosity, the Purdue team validated a popular nutritional hypothesis: that eating a small snack shortly before dinner might “prime” the pancreas to reduce post-meal blood sugar spikes. Their model shows support for that idea, suggesting that timing small snacks correctly could improve glucose control, much like a vaccine prepares the immune system for a subsequent challenge.

Toward truly personalized diabetes technology

Roughly 38 million Americans live with diabetes, and millions more manage blood sugar concerns while taking GLP-1 weight-loss medications. As wearable health tech expands, Purdue’s work offers a foundation for smarter, more personalized systems — including fully automated “set-and-forget” devices.

“This gives the community a complementary toolkit,” Fratus said. “Instead of trial-and-error, we can now design personalized patches informed by physics and physiology simultaneously, making the management of the diabetes as predictive as it is personal.”